Analýza chůze pomocí obrazů a moderních variant neuronových sítí

Abstract

This thesis focuses on the analysis of human gait from image data using keypoint detection. The aim was to design and implement an experimental application for processing motion data and comparing different approaches to gait phase detection. The proposed system combines the extraction of kinematic features with subsequent classification and enables systematic evaluation using both heuristic methods and deep learning models. The results show that neural networks achieve higher stability and robustness to noise in the input data, while heuristic approaches are more sensitive to the quality of keypoint detection. The thesis provides a modular experimental framework for human gait analysis and a foundation for further research in this area.